Comment on page
Stork for Low Latency dApps
What is Stork, and how to use it?
The Stork hybrid model maintains all the same guarantees of traditional decentralized "oracles": Stork data is fully verifiable, manipulation-resistant, and offered across decentralized infrastructure.
Hybrid Stork is an ultra low-latency (50ms), decentralized off and on-chain price feed designed any chain. Hybrid Stork is powered by a decentralized network of data publishers, and signed in a way compatible with signature verification on EVM, StarkEx, Starknet, and Move chains. We prioritize performance, using ultra-fast websockets, to ensure data is available at the millisecond level, similar to the data used for trading in TradFi.
Using an off-chain, decentralized feed gives protocols the ability to perform initial processing off-chain, and elect to only push on-chain the price updates that are relevant to your product. Since prices are signed in a chain-compatible way protocols can use Stork’s Verifier Contract, or their own smart contract, to verify the feed on-chain, proving that the data is legitimate.
To illustrate the cost and performance benefit of Stork, consider a traditional oracle operating on a performant L2 rollup like Arbitrum. A typical transaction can cost anywhere from 10-30¢, plus additional L1 calldata which can range from 0.1 - 1¢ per update depending on congestion and how much metadata is stored.
Let's say we wanted to receive a price update for a single market, every second, to match the L2's native blocktime. Assuming a 20¢ total transaction cost, the monthly cost can be calculated as follows:
That's around half a million dollars... per month. Scaling this up with additional markets, and allowing for multiple publishers for each price, significantly drives up the cost further.
For this reason, traditional oracles elect to push the price to chain selectively, for example every 30 minutes or during high volatility. Here, wee see Chainlink providing a price update every 30 minutes or when volatility exceeds 0.5%:
Chainlink XRP/USD Arbitrum Feed. Screenshot taken 3/9/2023
Meanwhile, using a hybrid protocol like Stork, where the protocol, liquidators, or other parties can monitor Stork's feed off-chain before submitting and verifying the prices on-chain, means that only a subset of price updates are needed. For example, if a liquidation happens once every 5 minutes (12 times per hour), and funding payments occur hourly, then using the same assumptions as above, the total cost to publish those prices on chain is:
Further, the transactions related to liquidating undercollateralized positions can be subsidized entirely with the fees already associated with liquidations.
Perhaps more importantly, the hybrid model allows for a more diverse feed set, so that prices can be customized to the appropriate use case. Traditional oracles are highly incentizied to only offer very popular feeds, in order to amortize their cost across as many users as possible. With a hybrid model, the data provider can make any arbitrary feed available, knowing that the cost is only incurred when the data is actually being used
- For Perpetual Exchanges, we recommend using the Perpetual Future Mark Price Index and/or the TWAP and Funding Fee Calculations.
- For settlement use cases (e.g. cash-settled options or futures), consider Stork's Consolidated Order Book Benchmark (COBB).
- For other use cases, contact us to define an appropriate feed that meets your latency, decentralization, and formality requirements.
Last modified 8mo ago